Journal of Materials Research and Technology 33 (2024) 5922–5933

Contents lists available at ScienceDirect

Journal of Materials Research and Technology

journal homepage: www.elsevier.com/locate/jmrt

Influence of adjustable ring modes on molten pool behavior and

microstructure of Al–Si coated boron steel in fiber laser welding
Hyun-Uk Jun a,b, Jooyong Cheon b, Yeong-Do Park c , Yang-Do Kim a,** , Changwook Ji b,*
a
Department of Materials Science and Engineering, Pusan National University, Busan, 46241, Republic of Korea
b
Smart Forming Process R&D Group, Korea Institute of Industrial Technology, Ulsan, 44413, Republic of Korea
c
Department of Advanced Materials Engineering, Dong-Eui University, Busan, 47227, Republic of Korea

A R T I C L E I N F O A B S T R A C T

Handling editor: P Rios The influence of adjustable ring modes on the phase transition and microstructural evolution in conduction mode
laser welding of Al–Si coated boron steel was investigated. Distinct geometric variations and molten pool
Keywords: behavior were observed through a ring-shaped beam effect. The ring beam introduced by ring-beam-mode (RBM)
Adjustable ring mode and dual-beam-mode (DBM) laser welding inhibited Al segregation into the fusion zone (FZ). The ring beam
Microstructural evolution
caused the Al molten particles to be driven towards the outer boundaries of the weld pool. A ring-beam-mode
Al–Si coated
(RBM) and a dual-beam-mode (DBM) laser welding inhibited Al segregation into the FZ, resulting from a min­
Ring-shaped beam
Al segregation imal temperature gradient and extension of the solidification time. Phase transition calculations demonstrated
δ-ferrite formation that a mount of residual ferrite within the FZ is significantly correlated with the Al concentration at specific
temperature points, indicating a predominant role of a central-beam-mode (CBM) FZ in δ-ferrite formation. The
RBM and DBM effectively suppressed Al segregation into the FZ, leading to a reduction in ferrite formation and
the development of a refined microstructure characterized by a uniform and high density of geometrically
necessary dislocations. Electron backscattered diffraction (EBSD) analysis revealed a higher proportion of low-
angle grain boundaries in the RBM and DBM, attributed to the unique thermal distribution induced by the
ring beam. These findings highlight the critical role of beam mode in tailoring the microstructure and properties
of laser welds in Al–Si coated boron steel.

1. Introduction This significantly reduces welding quality and makes it difficult to

achieve excellent mechanical properties.
High-strength steel (HSS) is becoming increasingly prominent in the Martin et al. [6]. investigated the austenite formation process during
automotive industry as a material that meets the demands for light­ the hot-press forming of low carbon manganese steels with varying Al
weight and high durability to address energy shortages and environ­ content. They demonstrated that aluminum acts as a strong ferrite sta­
mental pollution [1]. As cold forming alone cannot produce bilizer in the microstructural evolution. Wang et al. [7]. established a
complex-shaped parts, hot stamping processes that operate at austeni­ mechanism where the Al–Si coating on 22MnB5 steel dissolves in the
tizing temperatures between 850 and 950 ◦ C have been developed [2,3]. molten pool during laser welding, leading to the formation of δ-ferrite
This method is applied to boron steel with excellent hardenability, and phases in Al-rich areas. They also found that welds in argon atmospheres
the ultimate tensile strength can be achieved up to 2 GPa. To prevent are more susceptible to δ-ferrite formation compared to those in air. Sun
surface oxidation and decarburization of steel at high temperatures, the et al. [8] reported an increased presence of primary δ-ferrite within the
Al–Si coating with excellent high-temperature oxidation resistance and Al-enriched molten pool of welded joints, despite inconclusive results of
anticorrosion ability are widely used [4]. However, the Al–Si coating on improved tensile shear strength attributed to the Al–Si coating. Saha
the steel surface is mixed into the fusion zone (FZ) during laser welding, et al. [9] employed nanoindentation and TEM techniques to examine the
reducing the stability of austenite and forming the ductile ferrite phase, FZ of microstructure and hardness post hot-stamping. Their comparison
resulting in expanding the dual-phase (martensite/ferrite) region [5]. of α-ferrite and martensite nano-hardness substantiated the influence of

* Corresponding author.
** Corresponding author.
E-mail addresses: [email protected] (Y.-D. Kim), [email protected] (C. Ji).

https://doi.org/10.1016/j.jmrt.2024.10.244
Received 27 August 2024; Received in revised form 22 October 2024; Accepted 29 October 2024
Available online 30 October 2024
2238-7854/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

ferrite formation on the mechanical behavior of the welds. Laser welds operating in adjustable ring mode. The laser, with a maximum 6 kW
exhibit a microstructure composed of martensite and δ-ferrite, whereas power evenly distributed across its central and ring components,
TWB joints uniformly contain transformed α-ferrite originating from emitted two 1064 nm beams through a fiber optic cable featuring a 70
martensite and δ-ferrite. The elevated Al and Si content within the μm inner core and a 180 μm outer ring. A HIGHYAG BIMO welding head
α-ferrite may hinder martensite formation and promote ferrite growth. integrated with a 6-axis KUKA robot system provided mobility. The
They studied the relationship between microstructure and mechanical welding head incorporated a 250 mm focal length lens producing a 378
properties to quantify the impact of the Al–Si coating. μm focus spot. A 15 μm gap between two laser beams was filled with a
Laser welding is extensively utilized in the automotive, shipbuilding, low-refractive-index glass material. The ARM laser, equipped with
and aerospace sectors, benefiting from its rapid processing speed and independently adjustable concentric beams, facilitated three laser beam
limited heat affected zone (HAZ), this welding method excels in various modes: central beam mode (CBM), ring beam mode (RBM), and dual
applications. Compared to conventional laser welding methods, the beam mode (DBM). A Precitec YW50 laser head, containing a 150 mm
ARM laser, with its adjustable ring mode, provides greater flexibility in collimated lens and a 500 mm focal lens, was employed. Fiber laser
controlling energy distribution. This capability allows for fine-tuning of beams were configured in CBM, RBM, and DBM modes through external
the microstructure and mechanical properties within the FZ. Rasch et al. program adjustments of the central and ring beam outputs. Fig. 3 dis­
[10] investigated the welding characteristics of various shaped laser plays schematics of laser beam profiles at the optimal focusing. To
beams, demonstrating that a ring-shaped beam was most effective in examine the impact of laser beam profile and power distribution, the
stabilizing the weld pool. Similarly, Duocastella and Arnold [11] found laser power in the core and ring regions was varied. The core and ring
that mitigates thermal stresses and promotes a more controlled molten laser beams exhibited diameters of 147 μm and 378 μm, respectively, at
pool geometry, leading to improved weld quality by creating a uniform the focal point. Laser power for the core and ring fibers was adjusted
thermal gradient, the ring-shaped beam. Although the ring-shaped laser between 0 and 2000 W, and 0 and 3000 W, respectively. Weld condi­
offers advantages like reduced spatter and improved core quality, its tions are detailed in Table 2, with laser power distribution accounting
underlying weld mechanisms and process optimization remain areas for for energy density changes due to beam mode variations.
further investigation.
In this study, the influence of adjustable ring modes on the molten 2.3. Analysis methods
pool behavior and microstructural evolution in the welds of 1.2 mm
thick Al–Si coated boron steel using conduction mode laser welding Metallographic samples underwent polishing and subsequent
were examined. Three beam modes were selected as weld conditions by etching in a 2% Nital solution (98 mL C2H5OH + 2 mL HNO3) for 15–20
independently adjusting the central beam and ring beam. To clearly s. Microstructural analysis was performed through 3D optical micro­
understand the effect of adjustable ring modes, the molten pool behavior scopy (OM, HIROX RH-2000) and field-emission scanning electron mi­
was observed interlocking the laser welding machine with the high- croscopy (FE-SEM, Hitachi S-4800). The chemical composition was
speed camera. The trends of segregation and phase transformation ac­ determined using energy-dispersive X-ray spectroscopy (EDS, Oxford
cording to the molten pool mixing location of Al–Si coating layer, were Instruments). Phase transformations in the welding pool were analyzed
compared and analyzed via SEM and EBSD analysis. using JMatPro software. Fig. 4 illustrates the experimental set-up used to
capture high-speed images of the molten pool behavior during laser
2. Experimental procedure welding. A Photron Mini UX camera with a Micro-Nikkor 105 mm lens
was utilized to record images at a frame rate of 4000 frames per second.
2.1. Experimental materials The image was captured at a size of 1280 × 1024 pixels and a 12-bit
pixel depth. Fig. 4(a) shows an image set-up by interlocking a robot
1.2 mm thickness Al–Si coated 22MnB5 steel sheets were used in this with a one-axis sliding jig, and includes a moving workpiece for bead on
study, the coating layer and base metal (BM) chemical compositions are plate (BOP) test, and a stationary ARM laser beam and high-speed
listed in Table 1. A uniform 25 μm thick Al-10 wt% Si coating was camera as shown in Fig. 4(b). To further elucidate the microstructure,
applied to both sides of the substrate. Fig. 1 presents the microstructures EBSD analysis was conducted. Samples were tilted 70◦ , maintained at a
and line scanning analysis results of the BM and Al–Si coating. There was working distance of 15 mm and subjected to an acceleration voltage of
a Fe2SiAl7 inhibition layer (5~7 μm) between the steel substrate and 20 kV. EBSD maps were acquired with a 0.18 μm step size across a 65 μm
Al–Si coating as shown in Fig. 1(a). The BM, classified as a low-carbon by 49 μm field of view.
steel, consists of a ferrite matrix with pearlite and a small amount of
martensite as shown in Fig. 1(b). Additionally the electron probe micro- 3. Results and discussion
analysis (EPMA) mapping of the Al–Si coated layer is depicted in Fig. 2.
Considering the composition of the coating layer and the BM, the EPMA 3.1. Comparison of weld morphology for different beam modes
mapping was conducted for Fe, Al, and Si elements. The imaging was
performed under analysis conditions of 15 keV, 100 nA, and a resolution 3.1.1. Weld surface & cross-section
of 192 × 144. High-resolution microscopy revealed some Fe2SiAl7 of In laser welding, the formation of the molten zone generally involves
island-type and thin interlayer of Fe2Al5 and FeAl3, less than 1 μm two types: the keyhole type, where the bead is formed by melting and
thickness, between the inhibition layer and the steel substrate. the conduction type, as well as the evaporation of molten metal, where
the bead is formed by melting the adjacent solid due to thermal con­
2.2. Laser welding experiment duction of the molten phase [12]. The molten zone formed by the
keyhole phenomenon involves considerable stirring between the liquid
Experiments were performed using a HighLight FL-ARM fiber laser phase and the coating layer, resulting in mixing of the base metal and
coating layer components. In contrast, the molten boundary formed by
thermal conduction occurs with almost no stirring, forming a liquid
Table 1
phase. Fig. 5 shows the weld bead surface morphology of ARM laser
Chemical compositions of the Al–Si coated 22MnB5 steel sheet.
welds with different beam modes. Frist, all weld specimens are
Material Chemical composition (wt.%)
bowl-shape, resulted from thermal conduction mode. The welds of the
22MnB5 (1.2 C Si Mn Cr B Ti Al Fe RBM and DBM with ring beams showed traces of molten metal being
mm) 0.23 0.24 1.20 0.16 0.002 0.03 0.03 Bal. pushed to both sides of the weld surface. Moreover, the welds adjustable
Al–Si coating 10 90
ring mode had a large deviation of the weld surface along with undercut.
– – – – – –

5923
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

Fig. 1. FE-SEM micrograph and line scanning (a) of the Al–Si coating layer and the microstructure (b) of base metal.

Fig. 2. EPMA maps showing the elemental distribution of Fe, Al and Si element in the Al–Si coating on 22MnB5 steel.

Fig. 3. Schematic view of ARM laser beam profiling at optimum focal position used in the experiment.

finer than those of other beam modes, which was related to the wider
Table 2
weld pool melt and solidification time. On the other hand, ripple shape
Weld conditions in ARM laser welding by the conduction mode.
was not observed on the weld surface of the DBM, which was attributed
Beam Laser power Travel Laser beam Focal Irradiation to the minimum temperature gradient caused by maximized pre-heating
mode (kW) speed diameter length angle (◦ )
effect due to the application of two beams [11,13].
(m/min) (μm) (mm)
Center Ring Fig. 6 shows the weld bead cross-section of BOP samples for CBM
CBM 2 0 5 378 250 90 (Fig. 6(a)), RBM (Fig. 6(b)), and DBM (Fig. 6(c)). First, although the
RBM 0 3 same cross-sectional shape was shown for all laser welded specimens, a
DBM 1 2 difference in the region of the FZ and heat-affected zone (HAZ) occurred
at similar heat inputs. This was considered to be due to the temperature
The distribution of molten slag from the Al–Si coating on the bead gradient of the ring beam, as shown in Fig. 5, even though the laser was
surface was found to be contingent upon the laser beam mode, with irradiated at similar heat inputs considering the energy density distri­
accumulation occurring at either the weld edges or the centerline. The bution. Additionally, the distribution of white phases within the FZ for
single beam laser welds showed a typical characteristic of stable ripples, the dilution of the Al–Si coating was different for each beam mode. The
as shown in Fig. 5(a) and (b). The ripples of the CBM were relatively reason for the formation Fig. 6(d) shows the weld bead geometry of BOP

5924
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

Fig. 4. (a) Set-up experimental process for high-speed imaging, (b) schematic image of the ARM laser welding to analyze molten pool behavior by each beam mode.

Fig. 5. Weld bead morphology of weld surface cross-section at different parameters: (a) CBM, (b) RBM and (c) DBM.

samples for each beam mode using the image processing software considering the Al-10 wt%Si coating thickness (25 μm), the density
ImageJ. The geometry on the weld depth direction was dominant in the of coating (2.7 g/cm3) and the density of BM (7.8 g/cm3). The FZ for
CBM, while the geometry on weld width direction was dominant in the different laser beam modes are summarized in Table 3. The calculated
FZ of the RBM and DBM. The chemical composition, particularly the Al weld area of the FZ increased in order of CBM, RBM, and DBM, which
concentration (w(Al)%)), was calculated based on the weld geometry, was closely related to the distribution of energy density for different
including weld width (W) and area (S). The calculation was performed beam modes. As the weld area increased, the removed coating layer (CL)
using equation (1) [14]: also increased, so there was no significant difference in the Al content
inside the FZ. However, the distribution of white phases formed inside
W ⋅ tCL ⋅ ρAl ⋅ 0.9
w(Al) = × 100 (1) for each beam mode was observed to be significantly different. In the
S ⋅ ρFe
CBM, many large and small ferrite-separated regions were investigated

5925
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

Fig. 6. Weld bead cross-section at different parameters: (a) CBM, (b) RBM and (c) DBM, and (d) weld bead geometry graph.

steel pool and floats because surface tension has a positive temperature
Table 3
coefficient. They found that the configuration of the weld is directly
Comparison of area ratio of the diluted Al–Si coated layer (CL) in fusion zone
influenced by the fluid dynamics within the molten pool, with insuffi­
(FZ) for each beam mode.
cient backflow at the weld pool boundary leading to undercut. In the
case of the CBM, the molten Al–Si coating was observed to rapidly swell
and dilute into the laser beam irradiated area. In contrast, Al molten
particles in the RBM and DBM tended to rotate at the top, merge with
each other, and be pushed-out to the edge. In addition, the fluids was not
detected the molten pool back-flow, accordance with the results shown
in Fig. 5. As a result, the reduction attributed to Al segregation was less
Beam mode Area (x103 μm2) CL/FZ (%) Wt.% (Al) significant than that caused by temperature variations [21,22], so the
FZ CL fluids in the CBM flowed downward and were mixed with Al–Si coating.
Dissolved Al molten of the RBM and DBM, was pushed away without
CBM 427.84 41.80 9.77 3.04
RBM 578.19 55.47 9.59 2.99 being segregated in the melt zone due to the minimization of the tem­
DBM 674.19 64.47 9.56 2.98 perature gradient by the ring beam [13,23].
Furthermore, Fig. 8 shows the 3D constructive morphology to
observe the change in the edge between BM and the weld bead for weld
throughout the FZ because of the dilution of the Al–Si coating. In pool behavior by different beam modes. The edge of the CBM exhibited
contrast, δ-ferrite separation into FZ was significantly reduced despite the least variation as depicted in Fig. 8(a). The Al–Si coating was mixed
the increased Al–Si coating lost area in RBM and DBM. As a result, the within the FZ without significant displacement, resulting in a reduction
high Al content diluted into the CBM formed a significant proportion of in the coating layer thickness only within the weld bead area. In
metastable δ-ferrite phase within the FZ [15]. In addition, the contrast, significant height deviations were observed between BM and
pre-heating effect of adjustable ring mode extended the HAZ, which was the weld bead in the RBM (Fig. 8(b)) and DBM (Fig. 8(c)). The Al–Si
prominently observed in the DBM where two beams were simulta­ coating was displaced upwards beyond its original thickness in the BM
neously applied. The minimal thermal gradient in the FZ, attributed to area, consistent with the molten pool behavior presented in Fig. 7. As a
the slow solidification rate, facilitated a uniform distribution of the result, the Al–Si coating in the FZ showed a tendency to be pushed-out to
diluted Al–Si coating [13,16,17]. BM and stacked up in the RBM and DBM, and showed a more uniform
distribution in the DBM. The undercut occurred on the weld bead sur­
3.1.2. Analysis of molten pool behaviors face due to the unique energy distribution of the ring beam, resulting
In general, conduction mode laser welding does not involve intricate from the lack of back-flow at the molten pool boundary. Fig. 8(d) shows
physical phenomena, with the Marangoni force being the primary distribution graphs between BM and weld bead for each beam mode. In
driving force of the flow [18,19]. The high-speed images of the molten all three beam modes, the average height of Al–Si coating on BM area
pool behavior for various beam modes can be seen in Fig. 7. Independent was selected and compared as the reference point of the Z axis of the
control of the central and ring beams directly affects the thermal dis­ graph. Unlike the CBM, the RBM and DBM showed a deviation of about
tribution within the weld specimens, which changes the melting 80–85 μm due to the adjustable ring mode effect. The pushed-out Al–Si
behavior. Compared to CBM (Fig. 7(a)), a wider molten pool was pro­ coating from the weld bead area was calculated using an integral
duced by the adjustable ring mode lasers (Fig. 7(b) and (c)). Lin et al. equation as 22.49 and 27.72 μm2 in the RBM and DBM, respectively
[20] described that the first molten Al–Si coating separates from the

5926
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

Fig. 7. Molten pool behavior through high-speed camera imaging for each beam mode: (a) CBM, (b) RBM and (c) DBM.

Fig. 8. 3D constructive shape of the edge area between BM and weld bead with varied parameters: (a) CBM, (b) RBM and (c) DBM, as well as their X-Z axis dis­
tribution graphs.

[24]. Based on the calculations presented in Table 3, more than 40% of 3.2. Effect on microstructure and phase transition for different each beam
the coating was prevented from entering the FZ, effectively suppressing modes
ferrite formation and inducing microstructural changes within the FZ.
3.2.1. Al distribution and weld microstructure in the FZ
Fig. 9 shows weld microstructures of the FZ, fusion boundary, and
HAZ after ARM laser welding under different beam modes. The micro­
structure of single beam laser welds was susceptible to δ-ferrite

5927
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

Fig. 9. Weld microstructure for fusion zone and fusion boundary by each beam mode: (a) CBM, (b) RBM, and (c) DBM.

formation, with the CBM sample concentrated at the center of the FZ as (c), δ-ferrite formation in the molten pool decreased, indicating that the
shown in Fig. 9(a), and the RBM sample found mainly near the fusion ARM laser beam led to a more uniform Al concentration profile in the
boundaries as shown in Fig. 9(b) and in relatively small amounts. The FZ. In order to further investigate the influence of the Al content on the
observation of δ-ferrite in the FZ suggests that the Al–Si coating and evolution of microstructure, 10 points were scanned at various positions
Fe–Al IMCs were melted and diffused into the FZ. In contrast, the within the FZ and the average EDS results are presented in Table 4. The
microstructure of the DBM sample as shown in Fig. 9(c) was almost CBM FZ showed the highest average Al content (2.5 wt%), while the
entirely martensite stably formed at the FZ and fusion boundaries. Xu average Al content in the FZs with ring beam application clearly
et al. [25] attributed the intensified Al segregation to the limited decreased to 1.71 and 1.34 wt%, respectively. Since the Si content was
diffusion time and increased solute rejection under conditions of consistently low and exhibited minimal variations across all analyzed
reduced flow rate and accelerated solidification. This indicates a ho­ regions, as determined by EDS analysis, its impact on the microstruc­
mogeneous distribution of martensite throughout both the RBM and tural evolution was considered negligible for this study.
DBM samples because of the minimization of Al dilution into the FZ and Utilizing the determined chemical composition of the FZ, the
the stabilization of thermal distribution resulting from the adjustable
ring mode.
Table 4
The microstructures of the FZ for each ARM laser beam mode are
EDS results of average chemical compositions (wt.%) for different beam modes
exhibited in Fig. 10. SEM and EDS analyses revealed the presence of in the FZ.
δ-ferrite and lath martensite within the FZ. A mixture of ferrite and
Beam mode Fe Al Cr Mn Si
martensite is shown in Fig. 10(a), a large amount of δ-ferrite was formed
in the form of an irregular and long strip between martensite in the CBM 95.37 2.50 0.21 1.31 0.61
RBM 96.14 1.71 0.22 1.35 0.59
melted zone of the CBM sample, and the average Al wt.% of each phase
DBM 96.61 1.34 0.19 1.32 0.55
was shown. When the ring beam was added as shown in Fig. 10(b) and

Fig. 10. SEM & EDS analysis of the FZ at different parameters: (a) CBM, (b) RBM and (c) DBM.

5928
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

microstructural evolution for each beam mode was simulated using driving δ-ferrite formation at elevated temperatures. The primary
JMatPro software. The simulation results, presented in Fig. 11, show the δ-ferrite content in the CBM molten pool peaked at 90% during the
equilibrium phase transitions within the molten pool over a temperature initial stage I (L→δ). Despite subsequent transformations to austenite in
range of 600 ◦ C–1600 ◦ C, based on the average Al content of the FZ. both L+δ→γ and δ→γ stages, approximately 60% of the primary δ-ferrite
Dippenaar et al. [26] explored the relationship between varying Al persisted at 1400 ◦ C. Further δ-ferrite conversion to austenite occurred
content and austenite formation in low-carbon steel under continuous during the third stage, resulting in a notable decrease in δ-ferrite con­
heating conditions. They observed that Al element addition accelerated tent. Rapid cooling in stage III transformed austenite into martensite
ferrite transformation while suppressing austenite transformation, while retaining δ-ferrite, leading to a final microstructure of δ-ferrite and
attributing this behavior to a decrease in free energy change. The molten lath martensite in the FZ, accordance with Figs. 9 and 10. This is because
pool in CBM, most susceptible to Al dissolution, exhibited δ-ferrite phase very large δ-ferrite particles are formed during large undercooling in
formation and stabilization as shown in Fig. 11(a). Thus, the high laser welding, which limits the grain boundaries and multiple joints that
average Al content about 2.5 wt% was identified as the primary factor can be used as nucleation sites, making it difficult for austenite to

Fig. 11. Calculated phase transition diagram of the molten pool at the equilibrium condition by JMatPro software: (a) the FZ of CBM with 2.5 wt% Al, (b) the FZ of
RBM with 1.71 wt% Al and (c) the FZ of DBM with 1.34 wt% Al.

5929
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

nucleate [27,28]. Regarding the phase transformation depicted in energy [32]. Strain localization is commonly observed at martensite/­
Fig. 11(b) and (c) reveals a reduction in primary δ-ferrite content, from ferrite and martensite/martensite interfaces, a consequence of the
approximately 90%–80% after stage I, followed by a further significant relatively homogeneous deformation within individual phases [30,33].
decrease to less than 40% due to extensive peritectic reactions in stage II. The average KAM value of the CBM FZ was 1, which was the lowest
Upon completion of stage III at 1447 ◦ C, as shown in Fig. 11(c), a fully due to the dominance of δ-ferrite formation in the FZ. On the other hand,
austenitic structure was attained. Consequently, the effective segrega­ the average KAM values of the RBM and DBM samples showed high
tion control of Al within the DBM molten pool led to a near-full values of 1.12 and 1.30, respectively, because there was little aldilution
martensitic microstructure in the FZ following rapid cooling. In sum­ to FZ considering the results shown in Figs. 10 and 11. Saha et al. [9]
mary, the adjustable ring mode, through its enlargement of the melting explored the microstructural evolution within the FZ during laser
pool and extension of the solidification time, facilitated the diffusion of welding of Al–Si coated hot stamping boron steel. Their findings
the melted Al–Si coating, thereby enhancing the uniformity of Al dis­ revealed that dislocations present in the intitial δ-ferrite served as
tribution in the FZ. Therefore, the application of pre-heating in ARM nucleation sites for austenite. Thus, the δ-ferrite phase was devoid of
laser welding processes can be expected to diminish the formation of dislocations, rendering it more deformable. The defect-free ferrite phase
ferrite and augment the microstructural homogeneity in the FZ. within the FZ exhibited excellent ductility and a high capacity for energy
absorption. Typically, strain localization would occur at martensite/­
3.2.2. Evolution of microstructure via EBSD analysis ferrite or martensite/martensite interfaces, as deformation within each
In order to find out how the phase transformation affects the Al phase tended to be evenly distributed.
content, the microstructure of the FZ was examined by different beam The GND was estimated using equation (2) [34], derived from the
modes using EBSD analysis. Fig. 12 shows inverse pole figure (IPF) and average KAM value:
kernel average misorientation (KAM) maps, which were utilized to
2θKAM
characterize the microstructure for different beam modes. Typically, the ρGND ≅ (2)
bd
IPF map revealed the crystallographic orientations of ferrite and
martensite phases using the directional code triangle. In the FZ of the where ρGND is the GND density, θKAM is the average KAM angle (radian),
CBM sample, the δ-ferrite phase displayed a high degree of crystallo­ b is the Burgers vector magnitude (estimated as 0.248 nm for BCC iron),
graphic alignment with the neighboring martensite, as evidenced by the and d denotes the unit length (step size). Table 5 presents the calculated
IPF maps in Fig. 12(a). Fig. 12(b) and (c) presents the FZ of the RBM and GND densities for different laser beam modes. The CBM exhibited the
DBM samples, the limited formation of δ-ferrite promoted the direct lowest GND density in the FZ, whereas the DBM displayed the highest.
austenite-to-martensite transformation, resulting in a high degree of As a result, the DBM FZ contained the greatest amount of martensite
misorientation between martensite packets. among the three beam modes. δ-ferrite exhibits lower deformation en­
The KAM map indicates the average misorientation between a spe­ ergy compared to martensite. Martensite possesses higher deformation
cific point and its neighboring points within the same grain [29], and is energy due to a high density of GNDs and an alternating arrangement of
useful for assessing the dislocation density and stress state of the ma­ low-strength and low-hardness ferrite phases [13,35]. Therefore, the
terial. The KAM maps can be employed to quantify strain energy and absence of significant δ-ferrite in the RBM and DBM FZs, caused by the
GND density arising from phase transformations or rapid cooling [30]. lower Al content, is hypothesized to have induced a higher density of
To mitigate boundary effects, misorientation calculations were
restricted to intragranular neighboring points using a 5◦ threshold [29,
Table 5
31]. Fig. 12(d–f) reveals that regions with low KAM values (blue)
Geometrically necessary dislocation (GND) densities of the FZ area under
correspond to δ-ferrite, while those with high KAM values (green)
different laser beam modes.
represent martensite. The increase in KAM values is often attributed to
the accumulation of lattice defects, such as dislocations, which leads to a Beam mode CBM RBM DBM

higher degree of local lattice misorientation and increased stored elastic GND (m− 2) 5.80 × 1014 6.58 × 1014 7.08 × 1014

Fig. 12. EBSD results by different beam modes for IPF and KAM image: (a, d) CBM, (b, e) RBM and (c, f) DBM.

5930
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

GNDs within martensite particles, a consequence of the consistent shear 4. Conclusions

deformation during ARM laser welding.
Futhermore, Fig. 13 illustrates the microstructure of the FZ for This study investigated the effect of various beam modes on the
different beam modes using additional EBSD analysis. The grain molten pool behavior and the evolution of the microstructure in
boundary (GB) maps showed different distributions of martensite and 22MnB5 steel produced by conduction mode laser welding. The distri­
ferrite, as illustrated in Fig. 13(a–c). In this study, the grain boundaries bution of surface layers, microstructure, and Al segregation within the
with misorientation angle ranging from 1◦ to 15◦ were identified as low- FZ was examined for all beam modes. The dilution of the Al–Si coating,
angle grain boundaries (LAGBs). The CBM FZ exhibited a dual-phase phase transformation of the molten pool, and microstructural evolution
structure of martensite and ferrite, as shown in Fig. 13(a). This struc­ were analyzed using SEM & EDS analysis, thermodynamic simulations,
ture resulted in a high distribution of high-angle grain boundaries and EBSD analysis. The major conclusions could be summarized as
(HAGBs), marked by red lines, along the ferrite grain boundaries. In follows.
contrast, Fig. 13(b) and (c) showed a higher distribution of LAGBs,
marked by blue lines. The formation of non-equilibrium δ-ferrite during 1) The CBM mode showed a stable weld geometry with finer ripples,
laser welding is the origin of the characteristics of both the grain while the RBM and DBM modes introduced greater variations in weld
boundaries and the low-angle misorientation. The formation of non- bead shape and size due to laser beam interaction with molten pool.
equilibrium δ-ferrite during laser welding of 22MnB5 steel signifi­ Laser welding dispersed a strong Al element, resulting in a dual-
cantly influences the subsequent microstructural evolution, especially phase with long strip δ-ferrite.
the development of grain boundaries and low-angle misorientations. 2) The melting behavior of Al molten particles was pushed out of the FZ
The subsequent martensitic transformation, driven by the movement of using a ring beam, resulting in the formation and stability of δ-ferrite.
dislocations, is influenced by the presence of this metastable phase. The Al content and cooling rate strongly correlated with δ-ferrite
Fig. 13(d) and (e) further illustrate the misorientation distribution and formation, with the CBM FZ showing a higher tendency. The RBM
the statistical area fraction of LAGBs (1◦ < θ ≤ 15◦ ). The area fraction of and DBM FZs prevented Al–Si coating mixing and reduced δ-ferrite
LAGBs in the CBM was the lowest at 33.7%. In contrast, the area frac­ formation due to minimal thermal gradient and slow solidification.
tions of LAGBs in the RBM and DBM were approximately twice as high, 3) Phase transition calculations show that Al content strongly in­
at 69.8% and 74.2%, respectively. Previous studies have indicated that fluences residual ferrite content in the FZ, indicating that the CBM FZ
HAGBs in steel can inhibit martensite formation by obstructing the has a dominant effect on ferrite formation. The RBM and DBM
motion of lattice defects, commonly referred to as dislocations, which molten pools show less than 40% residual ferrite after the peritectic
are essential for martensitic transformation [36,37]. Consequently, the stage, compared to the CBM. Effective segregation control of Al
fraction of low-angle grain boundaries for different beam modes within the DBM molten pool results in a nearly fully martensitic
increased when ring beams were applied. The application of the microstructure in the FZ after rapid cooling.
adjustable ring mode led to an increase in the proportion of LAGBs 4) The EBSD analysis showed higher GND density in the FZ of RBM and
within the GB maps. This increase is due to the characteristics of the DBM compared to the CBM FZ, indicating higher lattice distortion.
energy density distribution of the ring beam, which pushed-out the Al–Si This is likely due to energy density distribution and minimal thermal
coating to the outer regions of the weld, preventing the formation of gradient with adjustable ring modes, resulting in high plastic
ferrite due to Al dilution. Additionally, the pre-heating effect of the ring deformation and low strain energy.
beam in the DBM case indicates that a more uniform distribution of 5) The microstructural evolution of δ-ferrite formation was significantly
martensite resulted from the stabilization of the temperature gradient influenced by adjustable ring modes, leading to the development of
[17,36]. LAGBs. EBSD analysis showed a higher proportion of LAGBs in the

Fig. 13. EBSD analysis of the FZ by beam mode for grain boundaries image: (a) CBM, (b) RBM and (c) DBM, as well as their (d) misorientation distributions and (e)
area fractions of low angle grain boundaries.

5931
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

RBM and DBM compared to the CBM, likely due to unique thermal surface application of colloidal graphite. J Mater Process Technol 2021;293:
117093. https://doi.org/10.1016/j.jmatprotec.2021.117093.
conditions induced by the ring beam. This increased fraction likely
[6] Martín DS, Palizdar Y, García-Mateo C, Cochrane RC, Brydson R, Scott AJ.
contributed to distinct microstructural evolutions. Influence of aluminum alloying and heating rate on austenite formation in low
carbon-manganese steels. Metall Mater Trans 2011;42:2591–608. https://doi.org/
10.1007/s11661-011-0692-1.
CRediT authorship contribution statement
[7] Wang XN, Chen XM, Sun Q, Di HS, Sun LN. Formation mechanism of δ-ferrite and
metallurgy reaction in molten pool during press-hardened steel laser welding.
Hyun-Uk Jun: Conceptualization, Investigation, Methodology, Mater Lett 2017;206:143–5. https://doi.org/10.1016/j.matlet.2017.07.008.
Visualization, Formal analysis, Writing – original draft, Writing – review [8] Sun Q, Di HS, Wang XN, Chen XM. Suppression of δ-ferrite formation on Al-Si
coated press-hardened steel during laser welding. Mater Lett 2019;245:106–9.
& editing. Jooyong Cheon: Software, Data curation, Resources. Yeong- https://doi.org/10.1016/j.matlet.2019.02.111.
Do Park: Writing – review & editing. Yang-Do Kim: Writing – review & [9] Saha DC, Biro E, Gerlich AP, Zhou NY. Fusion zone microstructure evolution of
editing, Supervision. Changwook Ji: Conceptualization, Writing – re­ fiber laser welded press-hardened steels. Scripta Mater 2016;121:18–22. https://
doi.org/10.1016/j.scriptamat.2016.04.032.
view & editing, Supervision. [10] Rasch M, Roider C, Kohl S, Strauß J, Maurer N, Nagulin KY, et al. Shaped laser
beam profiles for heat conduction welding of aluminum-copper alloys. Opt Laser
Data availability Eng 2019;115:179–89. https://doi.org/10.1016/j.optlaseng.2018.11.025.
[11] Duocastella M, Arnold CB. Bessel and annular beams for materials processing.
Laser Photon Rev 2012;6(5):607–21. https://doi.org/10.1002/lpor.201100031.
Data will be made available on reasonable request. [12] Ayoola WA, Suder WJ, Williams SW. Parameters controlling weld bead profile in
conduction laser welding. J Mater Process Technol 2017;249:522–30. https://doi.
org/10.1016/j.jmatprotec.2017.06.026.
Declaration of competing interest [13] Jun H-U, Kim J-D, Cheon J, Kim J, Park Y-D, Kim Y-D, et al. Microstructure and
mechanical properties of adjustable-ring-mode laser welded and hot-stamped Al–Si
coated boron steel joints. J Mater Res Technol 2024;28:844–55. https://doi.org/
The authors declare that they have no known competing financial
10.1016/j.jmrt.2023.12.018.
interests or personal relationships that could have appeared to influence [14] Lin W, Li F, Hua X. Effect of keyhole mode on weld shape and mechanical
the work reported in this paper. properties of thin sheets of Al-Si coated 22MnB5 steel. J Laser Appl 2018;30(4):
042001. https://doi.org/10.2351/1.5052155.
[15] Xu W, Tao W, Luo H, Yang S. Effect of welding speed on microstructure and
Acknowledgments mechanical behavior of laser welded Al-Si coated 22MnB5 steel. Opt Laser Technol
2022;154:108344. https://doi.org/10.1016/j.optlastec.2022.108344.
This research was supported by Korea Institute of Industrial Tech­ [16] Bardelcik A, Salisbury CP, Winkler S, Wells MA, Worswick MJ. Effect of cooling
rate on the high strain rate properties of boron steel. Int J Impact Eng 2010;37(6):
nology and the Ministry of Science and ICT (“Innovative hydrogen in­ 694–702. https://doi.org/10.1016/j.ijimpeng.2009.05.009.
dustry platform for carbon neutrality (One-stop platform for establishing [17] Chi J, Shao H, Song H, Wang X, Wu K, Zheng Q, et al. Effect of double-pulse
a hydrogen eco-system) (RS-2023-00304763)”) and the Ministry of frequency and post-weld heat treatment on microstructure and mechanical
properties of metal-inert gas welded Al–Mg–Si alloy joints. Mater Sci Eng A 2024;
Economy and Finance (MOEF) (“Development of Smart Manufacturing 913:147029. https://doi.org/10.1016/j.msea.2024.147029.
Technology for Low Temperature Fuel Tank for LNG Ships [18] Drezet JM, Mokadem S. Marangoni convection and fragmentation in LASER
(JA240007)”). treatment. Mater Sci Forum 2006;508:257–62. https://doi.org/10.4028/www.sci
entific.net/msf.508.257.
[19] Zhang R, Tang X, Xu L, Lu F, Cui H. Study of molten pool dynamics and porosity
Abbreviations formation mechanism in full penetration fiber laser welding of Al-alloy. Int J Heat
Mass Tran 2020;148:119089. https://doi.org/10.1016/j.
ijheatmasstransfer.2019.119089.
ARM adjustable ring mode [20] Lin W, Li F, Hua X, Zhang Z. Laser welding Al–Si coated hot stamping steel in
CBM central beam mode conduction mode: weld formation and Al-rich microstructure. J Mater Res Technol
DBM dual beam mode 2022;19:4282–95. https://doi.org/10.1016/j.jmrt.2022.07.012.
[21] Le TN, Lo YL. Effects of sulfur concentration and Marangoni convection on melt-
EBSD electron backscattered diffraction
pool formation in transition mode of selective laser melting process. Mater Des
EDS energy dispersive X-ray spectroscopy 2019;179. https://doi.org/10.1016/j.matdes.2019.107866.
EPMA electron probe micro-analysis [22] Lin W, Li F, Wu D, Chen X, Hua X, Pan H. Effect of Al-Si coating on weld
microstructure and properties of 22MnB5 steel joints for hot stamping. J Mater Eng
FZ fusion zone
Perform 2018;27:1825–36. https://doi.org/10.1007/s11665-018-3259-0.
GB maps grain boundaries maps [23] Liu Z, Wang X, Hang K. Enhancement of trapping efficiency by utilizing a hollow
GNDs geometrically necessary dislocations sinh-Gaussian beam. Sci Rep 2019;9:1–8. https://doi.org/10.1038/s41598-019-
HAGBs high-angle grain boundaries 46716-5.
[24] Jun HU, Kim JH, Kim JW, Lee EK, Kim Y Do, Ji C. Effect of electrode patterning on
HAZ heat-affected zone melting behavior and electrode degradation in resistance spot welding of A6014-
HSS high-strength steel T4 alloy. Korean Journal of Metals and Materials 2020;58:863–74. https://doi.org/
IPF maps inverse pole figure maps 10.3365/KJMM.2020.58.12.863.
[25] Xu W, Yang S, Tao W, Zhang J, Luo H. Effect of Al-Si coating removal state on
KAM maps kernel average misorientation maps microstructure and mechanical properties of laser welded 22MnB5 steel. J Mater
LAGBs low-angle grain boundaries Eng Perform 2023;32:4205–15. https://doi.org/10.1007/s11665-022-07401-3.
RBM ring beam mode [26] Dippenaar RJ, Phelan DJ. Delta-ferrite recovery structures in low-carbon steels.
Metall Mater Trans B 2003;34:495–501.
SEM scanning electron microscope [27] Zhou T, Zurob HS, Essadiqi E, Voyzelle B. Kinetics of delta-ferrite to austenite
phase transformation in a two-phase Fe-Al-C alloy. Metall Mater Trans A Phys
References Metall Mater Sci 2011;42:3349–57. https://doi.org/10.1007/s11661-011-0747-3.
[28] Moon S-C, Phelan D, Dippenaar R. New insights of the peritectic phase transition in
steel through in-situ measurement of thermal response in a high-temperature
[1] Xing ZW, Bao J, Yang YY. Numerical simulation of hot stamping of quenchable
confocal microscope. Mater Char 2021;172:110841.
boron steel. Materials Science and Engineering: A. 2009;499(1–2):28–31. https://
[29] Jia Q, Guo W, Wan Z, Peng Y, Zou G, Tian Z, et al. Microstructure and mechanical
doi.org/10.1016/j.msea.2007.09.102.
properties of laser welded dissimilar joints between QP and boron alloyed
[2] Xing ZW, Bao J, Yang YY. Numerical simulation of hot stamping of quenchable
martensitic steels. J Mater Process Technol 2018;259:58–67. https://doi.org/
boron steel. Materials Science and Engineering: A. 2009;499(1–2):28–31. https://
10.1016/j.jmatprotec.2018.04.020.
doi.org/10.1016/j.msea.2007.09.102.
[30] Calcagnotto M, Ponge D, Demir E, Raabe D. Orientation gradients and
[3] Nishibata T, Kojima N. Effect of quenching rate on hardness and microstructure of
geometrically necessary dislocations in ultrafine grained dual-phase steels studied
hot-stamped steel. J Alloys Compd 2013;577:S549–54. https://doi.org/10.1016/j.
by 2D and 3D EBSD. Materials Science and Engineering: A. 2010;527(10–11):
jallcom.2011.12.154.
2738–46. https://doi.org/10.1016/j.msea.2010.01.004.
[4] Ji C-W, Jo I, Lee H, Choi I-D, Do Kim Y, Park Y-D. Effects of surface coating on weld
[31] Echeverri EA, Nishikawa AS, Masoumi M, Pereira HB, Marulanda NG, Rossy AM,
growth of resistance spot-welded hot-stamped boron steels. J Mech Sci Technol
et al. In situ synchrotron X-ray diffraction and microstructural studies on cold and
2014;28:4761–9. https://doi.org/10.1007/s12206-014-1043-0.
hot stamping combined with quenching & partitioning processing for development
[5] Khan MS, Razmpoosh MH, Macwan A, Biro E, Zhou Y. Optimizing weld
morphology and mechanical properties of laser welded Al-Si coated 22MnB5 by

5932
H.-U. Jun et al. Journal of Materials Research and Technology 33 (2024) 5922–5933

of third-generation advanced high strength steels. Metals 2022;12(2):174. https:// [35] Sun Y, Wu L, Tan C, Zhou W, Chen B, Song X, et al. Influence of Al-Si coating on
doi.org/10.3390/met12020174. microstructure and mechanical properties of fiber laser welded 22MnB5 steel. Opt
[32] Shamsujjoha M. Evolution of microstructures, dislocation density and arrangement Laser Technol 2019;116:117–27. https://doi.org/10.1016/j.
during deformation of low carbon lath martensitic steels. Materials Science and optlastec.2019.03.024.
Engineering: A. 2020;776:139039. https://doi.org/10.1016/j.msea.2020.139039. [36] Chen H, Gu D, Dai D, Xia M, Ma C. A novel approach to direct preparation of
[33] Wang K, Wu M, Yan Z, Li D, Xin R, Liu Q. Microstructure evolution and static complete lath martensite microstructure in tool steel by selective laser melting.
recrystallization during hot rolling and annealing of an equiaxed-structure TC21 Mater Lett 2018;227:128–31. https://doi.org/10.1016/j.matlet.2018.05.042.
titanium alloy. J Alloys Compd 2018;752:14–22. https://doi.org/10.1016/j. [37] Sahoo G, Singh CD, Deepa M, Dhua SK, Saxena A. Recrystallization behavior and
jallcom.2018.04.148. texture of non-oriented electrical steels. Materials Science and Engineering: A.
[34] Moussa C, Bernacki M, Besnard R, Bozzolo N. About quantitative EBSD analysis of 2018;734:229–43. https://doi.org/10.1016/j.msea.2018.07.072.
deformation and recovery substructures in pure Tantalum. IOP Conf Ser Mater Sci
Eng 2015;89. https://doi.org/10.1088/1757-899X/89/1/012038.

5933